AI Resurfaces the Question: A Closed or Open Internet?
I’ve argued that 2025 is the most exciting moment in advertising, media, and advertising technology in a generation. There’s one reason: AI. And the adtech startups vying to lead the next era of our industry are releasing and growing interesting products in that space.
My client Firsthand, led by former FreeWheel and AppNexus leaders Jon Heller and Michael Rubenstein, announced a $26M Series A this week to scale its Brand Agent platform. In the coming weeks, at fellow Sharp Pen client Scope’s Landscape event and at the Markecture Live event, CEO and co-founder Brian O’Kelley will announce his company’s foray into AI. These sorts of innovations will transform how advertising works, disrupt the existing hierarchy of adtech solutions and, I believe, create billions of dollars in value.
The question is who will drive and own that value. Will the next fifteen years of adtech mimic the last fifteen in which a small number of companies (Google, Meta, Amazon, Microsoft, and Apple) controlled how consumers discovered products and information — and built massive advertising businesses due to their control of the flow of information? Or will AI actually decentralize consumer experiences, giving brands, retailers, and publishers the ability to set the terms for their own relationships with consumers?
We can think of the vectors along which the openness of the AI-driven internet will be decided as threefold: consumer utility, monetization, and data and technology. Let’s consider the cases for a more closed and open internet in the AI era along these three lines.
The Case for the Closed Internet
An internet dominated by a few walled gardens is what we’ve known for the past 15 to 20 years, and it tends to be easier to envision the persistence of the status quo than its disruption. There are a few reasons it’s likely the existing walled gardens or a set of new ones such as OpenAI will disproportionately orchestrate information and attention and dominate advertising and media.
1. Utility advantages: Everyone loves to criticize Big Tech, but they’ve created enormously valuable and scalable properties that have won consumer loyalty. Google has been the most valuable place to get information (a mantle it increasingly shares with the AI-native OpenAI). Meta is the most valuable place to connect with other people, see what people are talking about, and consume entertaining user-generated content. Amazon is the place to discover products. Apple is the portal to the entire app ecosystem. If these companies continue to get me the information, content, products, and digital services I want with unparalleled convenience, it’s not obvious why they should be disrupted as the primary portals to digital content, products, and services.
2. Monetization advantages: Utility allows Big Tech to generate massive audiences and attention. But they’ve also leveraged those advantages (their core cash cows) to build effective and increasingly automated advertising tools, and on top of those tools, they’ve added commerce to their ecosystems to close the loop (or vice versa in Amazon’s case), measuring and optimizing their ad engines’ ability to drive sales. In other words, the walled gardens aren’t just where you go to aggregate audiences; they are (or claim to be) the best places to convert audiences into revenue for marketers or audiences for publishers (buying sales for products or traffic for content). As long as Big Tech platforms continue leading the way on driving advertising performance at scale, they’ll only cement their dominant position in the media ecosystem.
3. Data and technology advantages: Winning begets winning. When you already have closed ecosystems that offer a range of use cases to consumers, massive audiences, and powerful audience monetization tools, you learn a ton about those audiences in the form of better and more scalable data, begetting better technology that maintains the audience and keeps the flywheel spinning. The Big Tech platforms have more capital, compute, data, and talent than smaller competitors. To be sure, competitors will emerge. But why the future landscape wouldn’t be dominated by a few companies with disproportionate financial, technological, informational, and human resources is hard to see.
The Case for the Open Internet
1. Better consumer experiences: AI won’t just fuel better experiences within walled gardens. It will fuel them on the open web, too. Firsthand, for example, is pioneering Brand Agents for marketers and publishers. Not only do these agents make shopping (to name one use case) on third-party sites more convenient (leveling the playing field with an aggregator like Amazon), but they also leverage a brand’s data and the context of an individual shopper’s experience to provide the best, most personalized possible version of shopping with that brand or retailer. No entity knows more about a merchant’s products than that particular merchant. So, in theory, the depth of this experience should provide a competitive edge that might encourage shoppers to go to independent sites instead of an aggregator. Ditto for publishers. True, you can get aggregated content on Meta’s platforms. But technologies like Firsthand’s Brand Agents may also allow publishers to create more powerful and interactive experiences on their own properties. Plus, AI-driven creative automation tools will lower the financial barriers to creating inventive content. So, the experience of the open web might become much more differentiated, personalized, and creative in the AI era, chipping away at the dominance of the walled gardens.
2. Better monetization: Over the last ten to 15 years, the open web has suffered from a scale problem. Why go to hundreds of different publishers to cobble together an audience when you could reach them all on Google and Meta? Moreover, even if you could cobble together that audience (for example, with the open programmatic market), signal loss meant that buyers could only purchase open web inventory with imperfect information, begetting fraud, illegal content, made-for-advertising sites, and the like. It seems possible to me that AI will solve this problem (and I know well-funded adtech companies are working on it). AI excels at processing and distilling highly complex data at massive scale. With better communication between the buy and sell sides, the next generation of DSPs (or whatever replaces them) may be able to facilitate a more perfect programmatic (or post-programmatic) ecosystem where buyers get scalable inventory and audiences — within their preferred brand parameters, tailored to the outcomes they desire, and without the unsavory elements of the open web. If that’s true, the open web will gain a monetization engine more powerful than programmatic 1.0 — and will better be able to compete with the walled gardens. The last component of this new open web equation will be the ability to drive, measure, and optimize performance, for which companies like tvScientific are already providing a model in the CTV space.
3. Democratization of technology: Allow me to leave it to OpenAI to summarize the bull case for the open web on this point: Open-Source AI can level the playing field. The past decade favored Big Tech’s proprietary algorithms and ad networks (Google’s search ranking, Meta’s feed algorithms). But in the AI era, open-source models (Mistral, LLaMA, Falcon) are improving fast, meaning independent companies can build AI-driven experiences without relying on Big Tech’s platforms. Open-source AI will power independent search engines, recommendation systems, and advertising networks, weakening the centralized advantage of Google and Meta. These innovations may allow independent companies (brands, retailers, publishers, and tech companies) to better compete with the giants.
The Vectors of a Closed vs Open Internet
In summary, we can think about the future openness of the internet as being decided by three vectors: the ability of the walled gardens and independent brands, retailers, and publishers to:
provide excellent consumer experiences
monetize their products, content, and audiences
innovate on technology and leverage data to succeed at the first two objectives
I could see these vectors pushing the information ecosystem in either a more closed or open direction. But at the very least, these are three of the factors that AI-era adtech companies should be thinking about.